1 Introduction

On February 6, 2023, two major earthquakes occurred in Turkey. As reported by the Ministry of Interior Disaster and Emergency Management Presidency (AFAD 2023), the first one was recorded at 4.17 am (01.17 GMT) in Kahramanmaraş- Pazarcık (epicenter coordinates: 37.2880 N–37.0430 E, Mw = 7.7, H = 8.6 km) on the NE-SW trending Eastern Anatolian Fault Zone (EAFZ), and the second one was recorded at 13.24 pm (10.24 GMT) in Kahramanmaş-Ekinözü (epicenter coordinates: 38.0890 N–37.239° E, Mw = 7.6, H = 7.0 km) on the D-W trending Çardak Fault (ÇF). These earthquakes occurred with a time difference of about 9 h and affected 11 provinces in a wide area of about 400 km from their epicenters. More than a hundred thousand buildings have been destroyed or severely and moderately damaged, and more than 50,000 people have lost their lives. In the studies prepared immediately after the earthquakes, it was stated that the total length of rupture for the first earthquake (Kahramanmaraş-Pazarcık earthquake) reached ~ 350 km, and the second earthquake (Kahramanmaraş-Ekinözü) reached ~ 160 km (Melgar et al. 2023; Goldberg et al. 2023). However, according to the strong ground motion records of the two events, field observations and information received from the people of the region, it was reported that the first earthquake was more active in Kahramanmaraş and Hatay, and the second earthquake was more active especially in Malatya (AFAD 2023). In the same report, the peak ground acceleration (PGA) of the first earthquake was recorded as 2039 cm/s2 (2.039 g) at Kahramanmaraş-Pazarcık station 4614, and the PGA of the second earthquake was recorded as 635 cm/s2 (0.635 g) at Kahramanmaraş-Göksun station no. 4612. In addition, in formal institutional reports prepared by AFAD (2023), Boğaziçi University, Kandilli Observatory and Earthquake Research Institute (KOERI 2023), Middle East Technical University (METU 2023) based on post-earthquake field observations, the main reasons why the earthquake was felt severely in such a wide area are summarized as follows: inadequate earthquake-resistant construction, lack of engineering projects in many buildings, wrong location choices for urbanization, and finally inadequate, incomplete and faulty earthquake-ground-structure relationship including the physical properties and behaviors of the earth. Moreover, several investigations reported that the main reason for the damage observed in the buildings was the inconsistency between the earthquake code requirements and the design/construction practices of the damaged buildings (Ozkula et al. 2023; Işık 2023). Especially, Ozkula et al. (2023) also specified that the concrete strength of the damaged buildings was quite low (i.e., 6–10 MPa).

Before starting any engineering construction, it is very important to evaluate the suitability of the site for the type of construction to be carried out. In this context, knowledge of the depth profiles and physical properties of the earth is necessary to understand the causes of damage after earthquakes as well as before earthquakes. Because, the most important causes of destruction and damage in earthquakes are the deformations that occur during the propagation of earthquake waves through any subsurface. Therefore, the damage that occurs anywhere after earthquakes due to the behavior of the ground increases the importance of knowing the soil properties in a sensitive manner. After many devastating earthquakes, geoscientific studies to investigate the causes of the damages have shown that the amplitudes of the earthquake waves and the dominant periods increase especially in weakly resistant soils and that liquefaction occurs in sandy soils where the groundwater level is shallower than 20 m. Moreover, it has been observed that earthquakes cause resonance between soil and structure, causing significant structural damage and loss of life even at distances of more than 100 km from the earthquake location (Tezcan and İpek 1973; Tezcan et al. 1978; Cassaro and Romero 1987; Kaptan and Tezcan 2012; Akkaya et al. 2015; Keçeli and Cevher 2015; Cevher and Keçeli 2018). The best examples of them are Russia-Alma Ata in 1964, Turkey-Gediz in 1970, Romania-Vrancea in 1977, Mexico City in 1985, Istanbul Gölcük in 1999, Van in 2011, Elazığ-Sivrice in 2019 and Aegean Sea-Sisam Island in 2020. In addition, it has been determined that the concave and convex structure of the underground bedrock topography causes earthquake waves to focus on the center of the basin and the edges of the basin, respectively, while sand lenses and dykes cause the earthquake waves to be channeled and therefore increase the amplitude of the waves (Keçeli 2012).

In the investigation of the physical properties and mechanical behavior of the ground, fast and noninvasive geophysical measurements have been commonly used in recent years, as opposed to traditional (drilling and open-pit) methods which are generally invasive, noneconomic and time consuming. Parameters obtained by conventional methods are static, whereas geophysical parameters are dynamic. Therefore, geophysical methods are considered the most reliable methods for seismic site characterization (Azwin et al. 2013; Sitharam et al. 2018) and are indispensable tools for seismologists and civil engineers to obtain sufficient information about subsurface structure and lithology (Benson and Yuhr 2002).

The depth profile of the subsurface allows the layered structure of the ground and the bedrock topography to be imaged. Besides, the physical properties and behavior of the ground include the determination of parameters such as Poisson's ratio, and modulus of shear, elasticity (Young's) and incompressibility (bulk) and the period of the soil vibration dominant, the soil amplification, the soil liquefaction and the available bearing capacity of geological units. Therefore, damages from earthquakes are directly related not only to structural characteristics but also to local geology and dynamic behavior of the soil. The seismic soil response at a given location is greatly influenced by the local geology and soil properties (Moustafa et al. 2007; Martínez-Pagán et al. 2018; Molina et al. 2018). This highlights the importance of site classification in areas of high seismicity or seismic vulnerability.

Shallow seismic refraction (SSR) and multichannel analysis of surface waves (MASWs) techniques are widely used to evaluate the basic design parameters of buildings and to explain soil and rock-related problems. Seismic refraction tomography (SRT) which is an inverse solution techniques of first arrival times from SSR data, and MASW techniques are time and cost effective in obtaining longitudinal wave (P-wave) and shear wave (S-wave) velocities of the subsurface materials, respectively. The SRT technique allows to obtain the 2D and 3D velocity structure of the subsurface (Rucker 2000; Leucci et al. 2006; Raghu Kanth and Iyengar 2007; Babacan et al. 2018), faults (Buddensick et al. 2008; Khalil and Hanafy 2016) and soil/rock parameters (Azwin et al. 2013; Keçeli 2012; Maraio et al. 2014; Sheehan et al. 2005; Vanlı Senkaya et al. 2020) without damaging the natural subsurface state. However, SRT technique is widely used in determining the interfaces between layers with different seismic velocities, engineering purposes, environmental projects, geotechnical investigations, dam safety and groundwater studies (Bridle 2006; Yılmaz et al. 2006; Hodgkinson and Brown 2005).

MASW is an active source method that is generally used to estimate the variation of S-wave velocity with depth at shallow depths (30 m), and its average for 30 m depth is called Vs30. It also helps to determine the engineering and elastic parameters of the soil. However, while the application of the MASW technique in the field is less time consuming, the processing of the recorded data requires high precision (Park et al. 1999). S-wave velocity (Vs) is calculated by inverse solution of the Rayleigh wave dispersion curve (phase velocity versus frequency) from the MASW data. Practically, the subsurface velocity-thickness model that gives the best fit between the measured and calculated dispersion curves is tried to be obtained (Xia et al. 1998, 1999; Miller et al. 1999).

Therefore, SRT and MASW techniques are used to estimate the physical properties of the soil materials and the engineering parameters required for the design of the structures to be built on it (Keçeli 2012; Kanli 2009; Yılmaz et al. 2009; Uhlemann et al. 2016). P- and S-wave velocities (Vp and Vs) are recognized to provide useful information for assessing liquefaction potential, natural vibration frequencies and the performance of ground motion (Bauer et al. 2001; Hunter et al. 1993; Tinsley and Fumal 1985). Hence, significant information about the stiffness of the soil is the key parameter for understanding the ground shaking response of soils and is highly effective for predicting soil amplification (Borcherdt 1994).

In recent years, Single Station Microtremor (MT) measurements have been utilized to determine local soil conditions. This method, developed by Nakamura (1989) and commonly referred to as horizontal (H)/vertical (V) spectral ratio (HVSR), is widely used as a fast, easy and highly reliable method for determining the dominant period and soil amplification characteristics (Lermo and Chavez-Garcia 1994). In the method, environmental noises (with frequency f <  ~ 1.0 Hz) are recorded as three components at a single station. A series of data processing operations such as detrending, filtering, windowing and smoothing are applied to the recorded data to obtain the HVSR curve, which is calculated as the ratio of the mean of the squares of the horizontal components to the vertical component. The frequency value (f0) corresponding to the maximum amplitude of the HVSR curve is determined as the soil dominant frequency (or period = 1/f0) value (Lachet and Bard 1994; Bard 1998). In addition, HVSR value is also used to determine the thickness of the soil layers on the bedrock by means of experimental relationships (Field and Jacop 1993; Birgören et al. 1998; Özalaybey et al. 2011). After Nakamura (1989), the method has been widely used by many researchers in engineering applications to analyses local soil conditions (Gallipoli and Mucciarelli 2009; Akkaya 2015; Akın and Sayıl 2016; Pamuk et al. 2017a-b; Öztürk et al. 2021).

The aim of this study is to obtain a preliminary knowledge about the relationship between the damages and physical properties of subsurface in four provinces including Kahramanmaraş, Hatay, Gaziantep, and Malatya after devastating earthquakes. Geophysical data were acquisited by authors on date from February 19 to 23, 2023, in the suitable locations in those provinces. In this context, SRT, MASW and MT measurements were made at eight different locations and two in each province, and then by analyzing the all data, physical, dynamic-elastic and geotechnical parameters of the locations were obtained. In addition, site classification based on Vs30 value was done according to TEBC code which is adapted from NEHRP. Furthermore, the information (epicenter distance, maximum peak ground acceleration-PGA, Vs30) for the strong ground motion station closest to the measurement profiles was evaluated and discussed in terms of damages and the subsurface characteristics obtained in this study. It should be noted that the data presented and conclusions drawn by the authors are the preliminary findings. Therefore, future researches involving more geophysical data and considering more PGA values will provide more comprehensive information about the structure, physical and geotechnical properties of subsurface.

2 Regional geology and seismotectonic

The locations of both earthquake and seismicity of the study area are shown in Fig. 1. Immediately after the main shocks, a large number of aftershocks are observed within a 15-day period, and this aftershock activity clearly increased the background seismicity of the region much more than before the main shocks. The region covering four provinces (Fig. 1), where geophysical measurements are carried out and where earthquakes are most effective, is within the influence area of the EAFZ, one of the most effective and active fault systems in Turkey, and is also very active in terms of current tectonics. EAFZ is approximately 580 km long (Duman and Emre 2013), NE-trending left-lateral strike-slip fault system consisting of many fault segments varying 50 km to 145 km (Duman and Emre 2013; Şahin and Öksüm 2021) that lies and between Karlıova triple junction point in Bingöl and Hatay Provincies (AFAD 2023; METU 2023). The EAFZ forms a major boundary between the Anatolian and Arabian Plates (Reilinger et al. 2006; Moreno et al. 2011), and the Arabian plate is moving toward the north-east with respect to the Anatolian plate at approximately 11 ± 2 mm/year according to GPS data (Çetin et al. 2003; AFAD 2023). Historically, many large earthquakes have occurred along the EAFZ (some of them are Antakya, 1822, Ms = 7.5; Karlıova, 1866, Ms = 7.2; Amik Lake, 1872, Ms = 7.2; Hazar, 1875, Ms = 7.1; Malatya, 1893, Ms = 7.1). = 7.1) occurred and caused loss of life and property (Kartal and Kadiroğlu 2013; Öztürk 2020). However, during the instrumental period, the devastating 1971 Bingöl (Ms = 6.5), 1975 Diyabakır-Lice (Mw = 6.6) and 2020 Elazığ-Sivrice (Mw = 6.5) earthquakes occurred.

Fig. 1
figure 1

Faults, earthquake locations, affected provinces and epicenter distributions of earthquakes (M ≥ 4.0) that occurred between February 6, 2023, and February 20, 2023 (modified from AFAD (2023)). Red stars and black triangles indicate the epicenter locations of the two earthquakes that occurred on February 6, 2023, and measurements locations in this study, respectively

Full size image

As reported in AFAD (2020), the region commonly contains Precambrian-Paleozoic and Mesozoic aged South-east Anatolian autochthonous or Arabian Platform sediments. On the platform sediments consisting of schist, gneiss, marble, limestone, sandstone, quartzite and shale rock units, there are complex units including allochthones ophiolite units, volcanic units and deep-sea sediments settled in the post-Upper Cretaceous region. The geomorphology of the region is shaped by the movements of faults, and it is seen that quaternary aged wide and thick alluvial materials (sand, clay, gravel and their alternations) with weak strength character, reaching up to 300 m, are stratified between the high mountains. However, sedimentary rocks dominate as the bedrock and the groundwater level is quite shallow (Palutoğlu and Şaşmaz 2017).

3 Methods and data analysis

Within the scope of this study, well-known and widely preferred geophysical methods were used to determine the geometric structure, physical properties and dynamic behaviors of the earth. To evaluate the basic parameters for engineering structures and to explain the problems related to subsurface materials (rock + soil), data were acquisited with SRT, MASW and MT methods, which are applied quickly and economically and evaluated in line with the principles of the methods. The SSR method includes the analysis of seismograms recorded in receivers placed in an order on the earth's surface, during the propagation of seismic energy sent into the ground from an artificial source, of directly incident waves and critically refracted waves at the interfaces of lithological units with different velocities. The first arrival times read from these seismograms are evaluated with the delay time (minus-plus and/or generalized reciprocal time method-GRM) and SRT techniques, and the seismic velocities of the lithological layers forming the ground structure (P-wave velocity, Vp and S-wave velocity, Vs), thickness and 2D velocity model are obtained. The SRT technique has a strong ability to characterize lateral and vertical velocity gradients in the study area. Moreover, it is ideal in environments with extreme topography or complex structures near the surface and where the operator has little or no prior knowledge of the subsurface structure (Zhu and McMechan 1989). Therefore, in any site characterization, the interpretation of SSR data with the SRT technique is highly effective compared to delay time techniques (Redpath 1973; Azwin et al. 2013). Therefore, SRT sections are widely used in determining the interfaces between layers, calculating the dynamic-elastic parameters of the ground, engineering and environmental projects, geotechnical research, determining the locations of covered faults, surveying dam sites and investigating the groundwater aquifer (Bridle 2006; Yılmaz et al. 2006; Lankston 1989).

The most important cause of damage occurring during an earthquake is the dynamic behavior of the ground during the passage of shear and surface waves. Shear stress, the basic parameter that determines the behavior of the ground under dynamic loads, is calculated from the shear wave velocity (Vs); therefore, this velocity value is extremely vital in the static project design of the structure. The most commonly and safely used method for determining Vs is MASW (Xia et al 1999; Park et al. 1999). In this method, the near-surface shear wave velocity structure is obtained as a function of depth (1D-Vs) by inverting the dispersion curve (change of phase velocity against frequency) of surface waves (Rayleigh and Love) recorded using an active source. Three basic steps of the method include data acquisited, data processing and inversion. Rayleigh type surface waves are analyzed when vertical component receivers are used in data acquisited. However, for Love type surface waves, horizontal component receivers are used. According to the 1D-Vs change, geotechnical characterization of the place, including the density-stiffness evaluation, soil-rock distinction and soil class, is performed. In particular, the Vs30 value is the key parameter used all over the world and taken into account in soil classification tables (NEHRP 2003; CEN 2005; TEBC 2018).

It is based on the principle of recording environmental vibrations (amplitudes varying between 0.1 and 1.0 microns and periods varying between 0.05 and 2 s) for a certain period of time (mostly > 30 min) in three components (East–West, North–South and Vertical). Since the single station MT method offers practical application opportunities, it has found widespread application in earthquake engineering applications and microzonation studies in recent years. The method is also known as the Nakamura (1989) HVSR method and is used to determine the ground dominant vibration period and ground amplification characteristics. In particular, it is used to obtain the resonance frequency, which is an important cause of collapse in earthquakes, and to estimate the bedrock depth using empirical formulas between frequency and thickness (Birgören et al. 2009; Zor et al. 2010; Özalaybey et al. 2011; Tun et al. 2016; Molnar et al. 2018; Buyuksarac et al. 2021).

3.1 Data acquisition

In this study, SSR, MASW and MT (approximately at the midpoint of the profiles) data were acquisited in eight profiles, two different profiles in each of the four provinces where the earthquake caused heavy damage. Six of the profiles were determined in places that exemplify the subsurface structure where there are buildings that were completely demolished or partly destroyed and damaged to different degrees. In these areas, since the streets and avenues are asphalt and concrete paved surfaces, geophones are placed on the ground with holder boxes. During data acquisited, especially, MT measurements were negatively affected by noise from working machinery, human mobility and traffic due to wreck removal activities in the surrounding area. One of the other two profiles is located at the place where the large landslide (splitting/opening) occurred in the olive grove in Tepehan Village of Hatay Province Altınözü District after the Pazarcık earthquake, and the other is on the surface trace of the fault line in Tevekkeli Village of Kahramanmaraş Dulkadiroğlu Municipality. Some images from the data acquisited locations are presented in Fig. 2.

Fig. 2
figure 2

Images from measurement locations: (a-1,2) Malayta-Yeşilyurt, (b-1,2) Hatay-Defne and Tepehan, (c-1,2) Gaziantep-İbrahimli and Nurdağı and (d-1,2) Kahramanmaraş-Tevekkeli and City Center. The slip of EAF in this place was observed as 4 m

Full size image

A 24-channel PASI 16S24-U model seismograph and a 4.5 Hz vertical component geophone, a 10 kg hammer and hardened plastic (chestamite) table with 20 cm diameter and 5 cm thick were used to acquisited the SR and MASW data. MT data were recorded for 30–45 min with a three component CMG 6-TD Güralp seismometer. Images of the equipment and connections used in seismic and microtremor data acquisition at the time of measurement are shown in Fig. 3.

Fig. 3
figure 3

Images of the measurement equipment during the field operation: a and b SR and MASW and c MT equipment and connection cables

Full size image

The field source-receiver layouts of SR and MASW data are given in Fig. 4, and the data acquisition parameters of all measurements are given in Tables 1 and 2. To gather the MASW data, the end-on (S1) and end-off (S7) shots in the SR profile were done, and only the recording time and sampling time were updated.

Fig. 4
figure 4

The layout of the source-receivers for SR and MASW data acquisition: S1,2,.. XS1,2… and XJ1,2,… show source points, source and geophone distance. dx and MT show geophone interval and location of the MT measurement, respectively

Full size image
Table 1 The data recorder parameters for SR, MASW and MT measurements
Full size table
Table 2 Soil classification according to Vs30 value in TEBC (2018)
Full size table

3.2 Data processing

All data obtained from each measurement method were evaluated according to the theoretical principles of the methods used. The 2D P-wave velocity model and geometry (layered structure, discontinuities and bedrock depth) of the subsurface were obtained from tomographic inversion of the first arrival times peaked from SSR data and will be referred to as seismic refraction tomography (SRT). One-dimensional shear wave velocity (1D-Vs)-depth profile was obtained from MASW data, and Vs30 value was calculated for soil classification. By applying a series of data processing to the three component MT data at each measurement point, the soil dominant vibration frequency (or period) and H/V ratio (HVSR) of the study area were determined. Analysis of SSR and MASW data was carried out with SeisImager/SW (2022), and analysis of MT data was implemented by Geopsy (2020) software.

An evaluation summary of SSR data is shown in Fig. 5 for an example data (Gaziantep-Nurdağı data) recorded within the scope of this study. First arrival times (direct and refracted waves) from each shot data are peaked (first arrivals for all shots are shown with green and red lines on the middle shot data in Fig. 5a) and plotted against receiver distances (blue lines in Fig. 5b). For an initial model, a 2D P-wave velocity model is obtained by iterative tomographic inversion (Fig. 5c). The solution starts with a homogeneous initial velocity model laterally; the solution is iterated until the root mean square (RMS) error between the measured time values and the calculated time values (black lines in Fig. 5b) is minimized. In our inversion analysis, generally, the RMS errors were below 5% within the 10 iterations. While the velocity-depth model (Fig. 5c) obtained when the minimum error is reached is interpreted physically according to lateral and vertical velocity changes, if there is drilling information about the profile, geophysical information can be interpreted geologically. However, since there is no drilling information in this profile, the interpretation was physically made only. Accordingly, the 2B-Vp section in Fig. 5c is interpreted as three layers including loose soils at first layer (Z =  ~ 0–5 m, 300–1000 m/s), changing from medium tight to very tight at the second layer (Z =  ~ 5–17 m, 1000–1800 m/s) and changing from soft to hard rock at the third layer (Z > 17 m, Vp > 1800 m/s). However, a covered normal fault anomaly was detected in the X = 5–20 m distance range of the section. In addition, it is seen that the bedrock depth decreases from Z =  ~ 20 m at X = 0 m to Z =  ~ 15 m at X = 62 m, that is, it becomes shallower.

Fig. 5
figure 5

The evaluation of the Gaziantep-Nurdağı (Gaziantep-2 in Table 4) SSR data: a picking on first arrival times on shot data, b the graph of the measured and calculated first arrival times and c 2D-Vp section obtained by SRT inversion and its interpretation

Full size image

The evaluation steps of MASW data are shown in Fig. 6 for the measurement profile in the Central Tevekkeli village of Kahramanmaraş Province. The steps include firstly the acquisition of the data from the field (Fig. 6a), imaging through frequency–phase velocity (f–c) transformation (Fig. 6b) and secondly peaking the fundamental mode dispersion curve (phase velocity values versus frequency) in the widest frequency range (Fig. 6c). Finally, Fig. 6d indicates the 1D-Vs depth profile by inverting this dispersion curve. The phase shift technique (Park et al. 1998) was used for the f–c transformation of the time record in Fig. 6a, and the damped least squares algorithm was used for the inversion process, with 20 iterations. Thus, by taking advantage of the change of the 1B-Vs depth profile up to a depth of 30 m, the Vs30 value, which is a very important and necessary parameter in the calculation of the dynamic-elastic and engineering parameters of the ground, geotechnical designs and earthquake hazard assessment studies, was calculated with the following formula.

$$V_{s} = \frac{30}{{\sum\nolimits_{i = 1}^{N} {\frac{{h_{i} }}{{V_{{s_{i} }} }}} }}$$

where, hi and Vsi present, respectively, thickness and shear wave velocity of the ith layer for 30 m depth, and N is the number of layers.

Fig. 6
figure 6

Evaluation stages of MASW data recorded in Kahramanmaraş Province Central Tevekkeli village: a MASW data, b f–c image, c Rayleigh wave dispersion curve and d 1D-Vs depth profile produced from the inversion of the dispersion curve. The Vs30 is calculated as 303 m/s, and this value represents ZD soil class according to TEBC

Full size image

By analyzing the MT data, the steps of determining the soil dominant vibration frequency f0 (or T0 = 1/f0) and H/V ratio are presented in Fig. 7 for three component (Vertical-V, North–South-NS and East–West-EW) data with 30 min recorded in Malatya-2 location. The data were processed with the open and freely available "Geopsy" program. In data processing, trending and a band-pass filter with a cutoff frequency of 0.5–20 Hz were applied to the raw data, respectively. Then, the spectra of all three components (Fig. 7b) were calculated for 25-s windows (Fig. 7a) (each colored line represents the spectrum of one window) in accordance with the SESAME criteria. These spectra were proportioned according to Nakamura's technique, and the H/V curve was obtained (Fig. 7c). By evaluating the H/V curve together with the spectra of each component (red frame area in b), the soil dominant frequency value (f0) on the H/V curve was determined (vertical gray bars in c). The black thick line in Fig. 7c shows the average of the spectra from each window, and the dashed lines show the standard deviation limits.

Fig. 7
figure 7

The analyzing of the MT data: a windowed of the three component MT data, b spectra for each component and c HVSR curve. In c, solid and dashed lines indicate HSVR curve and standard deviations, respectively. PSD in (b) means power spectral density

Full size image

4 Results and discussions

4.1 Evaluation and interpretation of profiles

Since there is no previously reported borehole information for measurement locations in this study, lithological soil and rock identification could not be done. In contrast, S-wave velocity information was attributed according to TEBC (2018) and Karslı et al. (2021). SRT sections and 1D-Vs depth profiles are given in Figs. 8 and 9. By correlating 2D-Vp velocity sections and 1D-Vs depth profiles, approximate velocity interfaces were determined, generally three layer subsurface models were created and the average P and S-wave velocities of each layer were obtained. However, using these seismic velocities, elastic (density, the modulus of Young's, Shear and Bulk) and geotechnical (soil classification in Table 2 according to Vs30 value, available bearing capacity, soil dominant vibration periods and amplification) parameters for each layer were calculated according to the formulas given in Table 3 and listed in Table 4. The soil dominant vibration periods were also calculated based on the Vs30 value empirically for comparison with those obtained from MT data.

Fig. 8
figure 8

2D-Vp velocity sections and 1D-Vs depth profiles: a,b Malatya-1, c,d Hatay-1, e, f Gaziantep-1 and g, h Kahramanmaraş-2. The dotted black curve in b, d, f, and h indicates Rayleigh wave phase velocity variation

Full size image
Fig. 9
figure 9

2D-Vp velocity sections and 1D-Vs depth profiles: a, b Malatya-2, c, d Hatay-2, e, f Gaziantep-2 and g, h Kahramanmaraş-1. The dotted black curve in b, d, f, and h indicates Rayleigh wave phase velocity variation. In g, considering the direction perpendicular to the plane of the screen or paper, the cross sign ( ⊗) and the dotted circle (ʘ) indicate a block moving away and a block approaching, respectively

Full size image
Table 3 Definition of elastic and geotechnical parameters used in this study
Full size table
Table 4 Elastic and geotechnical parameters at the profiles
Full size table

Even though they are located quite far from each other, SRT sections clearly present similar subsurface velocity models. In general, loose and less stiff soil units from the surface to a depth of ~ 7 m (Vp = 0.3–1.0 km/s, Vs = 0.16–0.3 km/s) and below that ~ 7–18 m thick (Vp = 1.0–1.8 km/s, Vs = 0.3–0.6 km/s) medium stiff-stiffer soil units are common. In the deeper part, the geological units with high weathering levels or soft rocks with Vp > 1.8 km/s, Vs > 0.6 km/s are placed. According to Vs30 values, the soils in the study areas are classified as ZD (Malatya-1, Hatay-1 and KMaraş-1) meaning the moderate stiff and ZC (Malatya-2, Hatay-2, Gaziantep-1, 2 and KMaraş-2) meaning fractured weak rock.

The ratio of P- and S-wave velocities, Vp/Vs, is a key parameter that provides information about the physical properties, lithological changes, mineral compositions, clay contents, grain sizes, fracture densities, porosity, fluid saturations and types of soil and rocks (Tatham 1982; Salem 2000; Uyanık 2011), and recently, it has been widely used for the characterization of soil and rocks (Keçeli 2012). According to Telford et al. (1976), Poisson's ratio is calculated based on the Vp/Vs ratio, and these two parameters are linearly related. Therefore, while the Poisson ratio approaches zero for very hard and solid rocks, it approaches 0.5 for very loose soils, highly weathered sediments and fluids (friable sediments or fluids). For moderate stiff materials, it is generally around ~ 0.25. The Vp/Vs and Poisson values vary between 2.19–4.43 and 0.37–0.47, respectively. However, weak materials and high porosity rocks have high Poisson's ratio and vice-versa. Accordingly, for the subsurface models obtained in all profiles, the first layer corresponds to weakly resistant, the second layer to moderately resistant and the third layer to resistant but possibly porous rock units.

The densities of lithological units were calculated with the empirical formula proposed by Uyanık and Çatlıoğlu (2015, 2019), which allows the combined use of P- and S-wave velocities and better represents low-velocity subsurface units. Thus, density values vary between 1.72 and 2.20 gr/cm3, indicating low-density subsurface units in the ~ 0–18 m depth range and partially dense sedimentary rocks at deeper depths.

In order to understand the level of resistance (endurance) of the geological units in the study areas against earthquake dynamic forces, thanks to seismic velocities, the elastic modulus including the shear (a measure of the solidity of the material), elasticity (or Young's, a measure of the hardness and strength of the material) and bulk (or incompressibility, a measure of compression under forces) modules is calculated and presented in Table 4.

When the obtained values are evaluated according to Bowles (1996), the geological units can be defined in three groups as: (1) loose (μ ≤ 600 kg/cm2, E ≤ 2000 kg/cm2), (2) moderately stiff and medium solid (600 < μ ≤ 3000 kg/cm2, 2000 < E ≤ 10,000 kg/cm2) and (3) strong to stronger (μ > 3000 kg/cm2, E > 10,000 kg/cm2). However, according to ASTM (1978), the incompressibility levels of these units are can be defined as low (400 < K ≤ 10,000 kg/cm2), medium (10,000 < K ≤ 40,000 kg/cm2) and high (40,000 < K ≤ 100,000 kg/cm2).

Allowable bearing capacity, Qa, is the load that the ground can carry without being exposed to any excessive deformation (collapse, settlement, sprain, shear fracture, shear, etc.) (Tezcan et al. 2009; Keçeli 2012; Uyanık and Gördesli 2013). The allowable bearing capacity values calculated for this study are lower than 10 kg/cm2 (~ 1 MPa). Values of Qa less than 4.0 kg/cm2 generally correspond to layers containing slightly to moderately compacted near-surface geological units, while values of Qa between 4.0 and 8.0 kg/cm2 represent layers containing more compacted soil units and mostly highly weathered rocks.

Soil amplification values calculated according to Midorikawa (1987) vary between 1.5 and 2.30. On the other hand, the ground dominant periods (T0 and MT0) calculated from Vs30 and directly determined by MT (HVSR) are generally consistent, except for Malatya-1 and 2 profiles, and vary between 0.12 and 0.40 s. When evaluated according to Ansal et al. (2001, 2004), these values indicate that the soil amplification and soil dominant vibration periods in the study areas are at low-medium levels. Moreover, for the soil dominant vibration period values according to Kanai and Tanaka (1961), it is defined as “rock-stiff sandy gravel units” and/or “alluvium consisting of sandy-gravelly stiff clay” and is in harmony with soil classifications based on velocities obtained from seismic profiles. In addition, the periods suggest that especially collapsed/seriously damaged 4–5 floor buildings are under the effect of resonance.

4.2 Possible relations between ground properties and damage observations

The findings of the present study provide more important clues that the damage levels that occurred after the earthquake and were directly observed in the field by the authors of this paper may be related to subsurface characteristics and behavior. Observations and relationships are summarized in Table 5.

Table 5 Damage observations and their relationship to ground properties
Full size table

Table 6 presents some information about strong ground motion (acceleration) stations located close to the measurement profiles. Information about the stations was obtained from AFAD. The selected stations are the closest stations that have both acceleration records for both earthquakes (Mw = 7.7 and Mw = 7.6) that occurred on February 6, 2023, and Vs30 values determined for the installation of the stations. Although the distances between measurement profiles and stations are less than 30 km, except for Malatya-1 and 2 profiles, other stations are very close to our measurement profiles. According to the epicenter distances in Table 6, the closest station to the epicenter locations of both the first and second earthquakes is station no. 4625 (R1 = 28.40 km and R2 = 65.22 km), while the farthest station is station no. 3136 (R1 = 148.38 km and R2 = 231.36 km). When the Vs30 values obtained previously for the installation of earthquake recording stations and the Vs30 values determined in this study are compared, except for the Vs30 value of station no. 4406 (Vs30 = 815 m/s), all the other values are in compliance with the Vs30 values obtained in this study, and inherently, it is seen that they are in the same soil class. The two closest measurement profiles (Malatya-1 and 2) to station no. 4406 are at a distance of ~ 28.0 km, and normally, the geological units within this distance may differ. Therefore, the Vs30 value (Vs30 = 815 m/s) at this station shows ZC soil class which means less weathered or moderate hard rock according to Table 2. However, the distance between Gaziantep-1 measurement profile location and station no. 2703 is ~ 3.5 km, and while the Vs30 value of the station is 758 m/s, the Vs30 value obtained in this study is 578 m/s. Normally, both values represent ZC soil class (360–760 m/s) as defined in Table 2. However, the value of Vs30 = 758 m/s is very close to the upper limit of the ZC soil class and therefore indicates a stiffer soil or less weathered rock than the measurement profile location (Vs30 = 578 m/s). Thus, this difference of ~ 180 m/s between both Vs30 values can be considered as a result of variability in local soil properties (degree of weathering, compactness, porosity, subsurface topography, etc.) at very close distances.

Table 6 Comparison of some information obtained from strong motion stations close to the measurement profiles in this study. R1, PGA1 and R2, PGA2 represent the epicenter distances of the first and second earthquake on February 6, 2023, respectively, and the maximum PGA values read at each station. Vs30_I and Vs30_P show the average shear wave velocities at 30 m depth obtained from the acceleration recording station locations and measurement profiles within the scope of this study, respectively. Stations parameters are taken from AFAD (https://tadas.afad.gov.tr)
Full size table

Figure 10a shows the distribution of maximum PGA values listed in Table 6 for the first and second earthquakes. The difference between the magnitudes of earthquakes is as ΔMw = 0.1. The energies of both events are calculated by using the formula, logE = 5.24 + 1.44Mw (URL-1 https://www.usgs.gov/programs/earthquake-hazards/earthquake-magnitude-energy-release-and-shaking-intensity (Accessed on November 17, 2023), Mw: moment magnitude). When the energies of both earthquakes are proportioned (E2 for Mw = 7.7 and E1 for Mw = 7.6), it is seen that E2/E1 ≅ 1.40. According to this difference, although the energy of the first earthquake was approximately 1.40 by times more than the second earthquake, it can be said that they produced similar amounts of energy. In this context, in the epicenter distance-PGA distribution in Fig. 10a, the accelerations of the first earthquake were less than PGA = 250 cm/s2 at stations no. 4406 and 2703, but larger at other stations, and it was quite high PGA = 638.32 cm/s2 at station no. 3124 which is 140.11 km away. On the other hand, the highest acceleration value of the second earthquake was observed at station no.4406 (467.20 cm/s2), while it was less than 100 cm/s2 at other stations. However, while acceleration values in general tend to decrease significantly with distance, it is seen that the acceleration values at stations 3124 and 3136 increased abnormally for the first earthquake. Nevertheless, it is understood that the epicenter distance-PGA change of the stations listed in Table 6 does not show a linear behavior. For example, station no. 3124 is 140.11 km away from the first earthquake location, PGA = 638.32 cm/s2, and 226.42 km away from the second earthquake location, PGA = 32.18 cm/s2. Similarly, station 3136 is 148.38 km away from the first earthquake and 236.36 km away from the second earthquake, and the peak acceleration values are PGA = 534.22 cm/s2 and PGA = 22.79 cm/s2, respectively. Although there is a distance of ~ 88 km between the epicenter distances of both earthquakes for station no. 3136, the dramatic decrease in PGA values of the second earthquake at the same stations is quite striking. Another interesting observation in Fig. 10a is that, although they are approximately at the same distance, the PGA values at stations 2703–8002-4625 and 4625–4406 are quite different from each other. It is clear that these differences cannot be explained only by epicenter distances and shallow subsurface characteristics of the station locations.

Fig. 10
figure 10

The PGA distribution and location of the stations: a variation of epicentral distance-PGA for nearest station locations in Table 6 and b locations of the station and profile according to fault directions. Faults are digitized from Melgar et al. (2023) in order to present their general directions. EAF: East Anatolian Fault, ÇF + SF: Çardak and Sürgü Faults and NF + PF: Nurdağı and Pazarcık Faults. M, G, H and K letters present profile locations in Malatya, Gaziantep, Hatay and Kahramanmaraş Provinces, respectively. The numbers of the ground motion stations closest to the survey profiles are indicated

Full size image

Considering the locations of the acceleration recording stations in Fig. 10b and the directions of the faults (or rupture direction) where the earthquakes occurred, it can be said that the direction of the fault and the PGA values are related. Namely, while station no. 2703 is in the zone approximately perpendicular to the direction of movement of NF + PF and EAF (first earthquake), stations no. 8002 and 4625 are located in the zone almost in the direction of the rupture direction. Similarly, while station 4425 is in the zone perpendicular to SF + ÇF (second earthquake), station 4406 is seen to be located in the same zone which is approximately parallel to the direction of these faults. Thus, these observations support that the magnitude of the accelerations and therefore the damages caused by earthquakes may be related not only to subsurface properties but also to the directivity effect in the propagation of earthquake waves (especially destructive surface waves).

5 Conclusions

The aim of this study was to investigate the soil character after two large earthquakes that occurred on February 6, 2023, and affected 11 provinces in the vicinity. For this reason, SRT, MASW and Microtremor measurements were carried out in eight different profiles in four provinces (Kahramanmaraş, Hatay, Malatya and Gaziantep) where the damage and loss of life after earthquakes were intense, and the first results obtained are presented. Also, the results obtained were correlated with the station information (epicentral distance, maximum PGA, Vs30) closest to the measurement profiles. SRT technique provided P-wave 2D velocity-depth models, MASW technique provided 1D-Vs depth profiles and MT (HVSR) technique provided soil dominant period values. Using the velocity information, physical and geotechnical parameters of the earth materials were calculated, and soil classification was made. All but two of the measurements (Hatay-2 (Tepehan) and Kahramanmaraş-1 (Tevekkeli)) were made in settlements. However, these two points are located very close to settlements (such as villages). According to the soil classification using Vs30 values, ZD (Malatya-1, Hatay-1 and Kahramanmaraş-1) and ZC (Malatya-2, Hatay-2, Gaziantep-1,2 and Kahramanmaraş-2) soil classes were determined. In all profiles, the subsurface model was determined as three-layered using SRT sections and evaluated together with the 1B-Vs depth profiles, and the P- and S-wave velocities of each layer were assigned as average. In general, the high velocity ratios (Vp/Vs > 2.5) and Poisson's ratios (σ > 0.35) explain that the near-surface geological units are very loose and the deeper rocks are very porous/fracture-cracked. The variation of dynamic-elastic parameters indicates weakly resistant geological units to depths of ~ 15–20 m in Malatya-1 (Yeşilyurt), Hatay (Defne) and Kahramanmaraş (Tevekkeli), and moderately hard geological units to depths of ~ 10 m in Malatya-2 (Yeşilyurt), Gaziantep (İbrahimli and Nurdağı), Hatay (Defne) and Kahramanmaraş (City center). However, in all profiles, the soil units close to the surface have very low available bearing capacities (less than 4 kg/cm2 or 0.4 MPa) and moderate (between 4–8 kg/cm2 or 0.4–0.8 MPa) toward deeper depths. Although there is no drilling data to define the lithology, all geophysical parameters show that the deep layers (usually Z ≥ 15 m) in the studied locations consist of porous, weathered sedimentary rocks, while above these units, mostly very loose soil materials (clay, silt, sand, gravel and their mixture) are present. However, regardless of the soil class, according to the station information close to the measurement profile, it has been observed that stations at close epicentral distances but within the zone parallel to the direction of movement of the faults have higher acceleration values than those in perpendicular zones. These observations indicate that there is a directivity effect in the propagation of earthquake waves.

Finally, all the findings of this study reinforce the view that the damages caused by earthquake-soil structure relationship was not fully and accurately reflected in the building designs, beyond the poor quality of construction materials, subsequent interventions to the buildings and structural defects. In order to improve the results obtained in this study, both the measurement profile and the number of stations to be considered should be increased, and geophysical measurement studies should be carried out to image deeper geological structures.